In about 10 years we could all feel a little closer, technologically speaking, to George Jetson. Although our cars probably won’t be flying, they could be driving themselves. In June 2008, Carnegie Mellon University issued a press release stating that their robotics institute had just received a $5 million grant from General Motors. This money will be used to continue one of the institute’s main projects: autonomous cars. According to Dr. Raj Rajkumar, who is the co-director of the project, the technology for driverless cars could be available to the public in as little as 10 years with the cars themselves available by 2030.
On November 3, 2007, DARPA sponsored a race between autonomous cars developed by 11 different groups. DARPA’s Grand Challenge website has stated that 89 teams applied to race, 35 passed DARPA’s initial tests and 11 were allowed to compete in the final event. “Boss”, an SUV entered by Rajkumar’s associates, Tartan Racing, finished first and won $2 million. The mission was for each car to not only cross the finish line but to also stop at pre-set checkpoints. Since half of the race was off-road, each car was given a GPS map of the 55-mile course 24 hours in advance. This gave scientists 24 hours to upload the map to the car. However, the positions of the checkpoints were only provided five minutes before the race began. Using 19 sensors, 2-D and 3-D lasers, cameras and short and long range radar, Boss was able to stop at each checkpoint and ultimately win the race.
Fortunately, Rajkumar believes a “hybrid or fully electric vehicle would be [an] ideal” platform to build autonomous cars upon since they have a built-in power source. This could be used to run the accoutrements used to drive the car. While Boss required an extra generator to power the various sensors, lasers and cameras, the new model is based on a Chevrolet Tahoe hybrid which extinguishes this need. Boss was also outfitted with bulky sensors and attachments around the top of the frame; future models are intended to have scaled down versions which can be hidden in the front and rear bumpers as well as the head and tail lights. Rajkumar expects some of the technology to be integrated into current cars in as little as five years. Cameras and lasers could be incorporated to sense when a driver is drifting off-road, if a car is hiding in a blind spot or to quickly slow down if a car ahead brakes suddenly.
Many social and financial benefits could come from the general acceptance of autonomous cars. For example, the elderly could be driven anywhere they wish without having to depend on others or their eyesight after dark. Children could automatically be driven to school with the push of a button, leaving parents free to go to work. Since cars would be controlled by a computer and not manually, accidents would decrease. According to the National Highway Traffic Safety Administration (NHTSA), there were almost 43,000 deaths from roughly 6 million car accidents in 2006. The National Center for Statistics and Analysis has reported that in 2005, car accidents were the leading cause of death for Americans ages 4 to 34. Rajkumar anticipates that the number of deaths could be lowered by 1 to 2 orders of magnitude to 430-4,300 deaths per year. This dramatic decrease would be due to the fact that computers do not get drunk, angry, fall asleep or commit human errors. The vehicle would always be fully aware. This would be positive for the overall economy since the NHTSA estimated in 2000 that $230.6 billion was spent every year due to car crashes. Of this, $977,000 was spent for each fatality and $1.1 million was spent on injured victims. If the number of accidents decrease, then it would follow that insurance rates would as well. Since insurance rates are based on how many accidents a type of car has, insurance for automated cars would have to be significantly less than for a manually operated one.
This is not to say that there may not be problems with such a car. Problems could arise if the car’s computer were “hacked” into or if the computer shorted out. If either of these were to happen, the consequences could be serious. Rajkumar believes the safety systems involved in autonomous cars would prevent either of these from being a problem. “Serious and secure safety mechanisms” would be built into the car to prevent others from accessing the computer. He doesn’t foresee problems being any worse than they are today. As for internal computer problems, there would be “multiple stages of operation.” If the computer were to go down, a backup would take over so fluidly, the driver would probably never know there was a problem. This is considered stage one. At stage two, the car would go into a more degraded operation or “limp home” mode. Stage three would allow the car to slow down, pull over to the side and call for help. The car itself would know if something was wrong and initiate the appropriate backup system.
Rajkumar believes GM will be one of the first manufacturers to begin production of these cars. “GM is our biggest partner and funding [organization],” he says. GM is even funding a second lab to continue such research to perfect the technology. However, anyone who watches the BBC’s top rated car show Top Gear will have already seen an autonomous car in action. Made by BMW, this experimental car flew around the racetrack while a nervous Jeremy Clarkson sat passively in the driver’s seat. Regardless of who produces the car first for the general populace, it is certain that this type of technology could shift how we think of driving and how much we pay individually and as a society for it.